How to re-import an atlas?

Hello all,

When starting my protocol, I imported the USCBrain with corresponding atlas. I am currently trying to compute the scout time series, but the atlas seems to have dissapeared (I fear I might have overwritten the anatomy folder when computing the FEM headmodel). My anatomy folder currently is the following:

image

I can import the USCBrain template again, but this would overwrite my current folder, which would delete my earlier work I assume.

When right-clicking the default anatomy, I can select the add MNI parcellation, but the USCBrain atlas is not present there.

Is there a way to re-import this atlas without overwriting my obtained results?

Thanks in advance!

I am currently trying to compute the scout time series, but the atlas seems to have disapeared

I'm not sure what you mean with this missing atlas...
What do you call "an atlas"?
What is missing exactly, and where was that information coming from?
Comparing with the screen capture from your previous post, it looks like you have more files than before.

Dear @Francois,

Thank you for the reply. I call it atlas, but I think what I mean is parcellation/scouts (I used the following tutorial to obtain the USCBrain atlas: https://neuroimage.usc.edu/brainstorm/Tutorials/DefaultAnatomy?highlight=(uscbrain)) and it contains 130 something scouts. What is "missing" currently are the scouts.

When I try to compute scout time series for an EEG file, I get the following message:

image

And when I look in the anatomy tab, I can't select any scouts (everything is greyed out, see below):

While previously all scouts of the USCBrain atlas were located here.

I hope is makes some sense,

Thanks in advance!

Thank you for the reply. I call it atlas, but I think what I mean is parcellation/scouts (I used the following tutorial to obtain the USCBrain atlas: https://neuroimage.usc.edu/brainstorm/Tutorials/DefaultAnatomy?highlight=(uscbrain)) and it contains 130 something scouts. What is "missing" currently are the scouts.

If you right-click on the anatomy > Use template > USCBrain, the cortex that is added to the folder contains the SVReg scouts.
If you re-process the MRI with SimNIBS, it may delete the existing cortex surfaces (with the SVReg atlas), but it would replace them with other cortex surfaces with the CAT12 atlases.

And when I look in the anatomy tab, I can't select any scouts (everything is greyed out, see below):

You need to have a cortex surface loaded and displayed in order to have menus available in this tab.

Dear @Francois,

Thanks for the quick reply. I think this indeed happened. When I click on the Use template -> USCBrain, I get the message that some files will be overwritten. Would this mean that I need to redo the SimNIBS calculations or is it not a problem since I already have the headmodel?

My apologies for the many questions.

Create a new subject with "Default anatomy:NO", set the USCBrain template for this subject.
Then you can copy-paste files between these two folders (for example copy the SimNIBS FEM model to the unmodified USCBrain template), as then should have the exact same reference and fiducial points (you need to double-check this before).

Dear @Francois,

Alright, so I made a new subject, which currently has the following anatomy folder:

image

For comparison, my default anatomy folder has the following structure:

image

I do not know exactly which files I should copy from the default anatomy to my Subject_USCBrain_Atlas, but I think more files need to be copied.

Furthermore does the functional data folder for my new subject look like this:

image

with DUNEuro FEM (volume) being the headmodel (obtained by following the tutorial: https://neuroimage.usc.edu/brainstorm/Tutorials/Duneuro)

When trying to compute the sources and subsequently extracting the scout time series I get the following error message:

Which I assume has something to do with the anatomy folder, but somehow talks about the default subject, which my current subject does not "inherits" from.

I'm fearing that my "Frankenstein" construction for extracting scout time series might be becoming too complex (given my initial attempt for running everything with the default subject and anatomy), which will hurt any possible replication attempts. Would you think restarting the whole process in a new protocol and importing the FEM models there to be the better option currently?

Yes, I definitely think you should create a new protocol and start again from scratch.
The risks of errors are too high...

Dear @Francois,

Thanks, I started this morning with the new protocol, the chance of error becomes too high indeed.

For clarrification: My main goal is to compute scout time series from resting-state EEG recordings. When these time series are calculated, I want to export them into MATLAB to perform my own analyses on them.

So as a recap then, I think I should do the following steps:

  1. Start a new protocol.

  2. Creat a new subject with NO default anatomy. For this subject I import the USCBrain template.

  3. Follow tutorial 20 (Headmodel; https://neuroimage.usc.edu/brainstorm/Tutorials/HeadModel)

2.1. Forward model: I calculate a FEM mesh from the T1 MRI image that comes with the USCBrain (I follow the DUNEuro tutorial: https://neuroimage.usc.edu/brainstorm/Tutorials/Duneuro)

  • Here I believe the initial problem started, if the FEM mesh calculation gives the popup "Overwrite data in the anatomy folder" or something similar, can I click "NO" and still save the results from the calculation?

2.2. Forward model: I have an EEG electrode file and adapted them so they are placed correctly on the head. I use this file for "compute head model". As option for the head model computation I use FEM forward model (See DUNEuro tutorial).

  1. Once I have the headmodel, I calculate the noise covariance matrices for my EEG data.

  2. For each EEG datafile then I calculate the sources ("Compute Sources 2018", with the choice of Minimum Norm and Current Density Maps, given my goal)

  3. I use the function "Scout time series" to obtain the timeseries of the scouts that are present in the USCBrain atlas.

Is my reasoning correct here?

Thanks in advance!

Dear @Francois,

As I thought, the FEM mesh generation is almost finished, and I get the following message:

image

so I just selected "No" and immediatly get the following error messages:

image

image

And I see no changes in my anatomy folder:

image

I think that I need to select "Yes" then, but then I get the problem from which this thread is started?

Any idea on how to fix this?

You pipeline looks good.

Here I believe the initial problem started, if the FEM mesh calculation gives the popup "Overwrite data in the anatomy folder"

I fixed these problems with the two commits below:

For each EEG datafile then I calculate the sources ("Compute Sources 2018", with the choice of Minimum Norm and Current Density Maps, given my goal)

If the different EEG files are imported in different folders: yes.
If they are in the same folder, you only need to compute one shared imaging kernel for all of them.